Development of a finer-resolution multi-year emission inventory for open biomass burning in Heilongjiang Province, China Authors: Bin Li, Zehua Xu, Baiyin Liu, Zhimiao Zhang, Wenting Qiu, Wei Wang Published: 2024 ABSTRACT Open biomass burning (OBB) is a significant source of air pollutants, profoundly impacting regional air quality and global climate change. However, due to the lack of high-resolution burned area products, limited biomass data resolution, and inappropriate emission factors (EFs), current OBB emission inventories have significant uncertainties. In this study, we integrated the FireCCI51 burned area product (250 m), high-resolution gridded biomass data, localized EFs, and various statistical survey data to compile a finer-resolution (250 m) and long-term (2001–2020) inventory of 11 pollutants for Heilongjiang Province (an important agricultural and forestry region in China). The results indicated that the annual average OBB emissions in Heilongjiang Province were 12.26, 63.70, 931.60, 17,203.35, 16.09, 171.25, 39.83, 71.22, 184.33, 129.16, and 6.35 Gg for BC, CH4, CO, CO2, NH3, NMVOC, NOx, OC, PM10, PM2.5, and SO2, respectively. Taking PM2.5 as an example, emissions from cropland, forest, and grassland burning accounted for 73%, 26%, and 1% respectively. Geographically, emissions were primarily concentrated in the western, southwestern, and northeastern regions of Heilongjiang. The peak PM2.5 emissions exhibited notable seasonal variation, with maxima occurring in April and October. This is primarily due to the extensive burning of straw during the spring and autumn ploughing seasons. The results of this study provide a detailed multi-year inventory crucial for atmospheric transport models and can support the development of effective pollution control strategies and greenhouse gas mitigation measures. Introduction Open biomass burning (OOB) is typically defined as the open combustion of vegetation resulting from human activities or natural causes. OBB emits significant quantities of greenhouse gases (CO2, CH4) and trace gases (CO, NOX), degrading local air quality and potentially contributing to global warming1,2. Furthermore, OBB generates hazardous fine particulate matter (PM2.5 and PM10) and other toxic volatile organic compounds, posing a substantial threat to human health3,4. The traditional method of allocating total emissions to grids based on population or cropland area weights has significant uncertainty5–7. The application of high-resolution satellite imagery has greatly enhanced the accuracy of monitoring and quantifying OBB activities and their emissions. The MODIS satellite product, characterized by its extensive temporal coverage and high spatial resolution (500 m), offers reliable data for researchers to monitor and evaluate global fires. Previous researchers have successfully used the MCD64A1 product to analyzed the spatial–temporal distribution characteristics of open biomass burning (OOB) in forest fires8. This kind of bottom-up approach to fire detection can substantially decrease uncertainty in emissions quantification. However, due to its limited resolution (500m), MCD64A1 products are unable to detect many small fires, leading to an underestimation of emissions. To reduce the omission of burned areas, some studies have integrated MCD64A1 products with MCD14ML products to enhance the completeness of burned area detection9. The MCD14ML product, based on active fire detection methods, has the capability to identify smaller fire-affected pixels (as low as 1/20 of a pixel). For instance, Li used the MCD64A1 product in combination with the MCD14ML product to study the inventory of polycyclic aromatic hydrocarbon emissions in the lower Yangtze River Delta, China10. The results suggested that this integration improved the accuracy of emission inventories compared to estimates based on single datasets. Nevertheless, this approach still exhibits certain constraints. The MCD14ML product can only collect fire during a specific time period (approximately 10:30 AM to 1:30 PM), thus providing limited information on the burned area11. Furthermore, the inherent 500 m resolution of the MCD64A1 product may result in missed fires. While combining these two products could yield additional burned area data, it may not be optimal. These limitations contribute to the underestimation of OOB emissions. Hence, higher-resolution and more precise burned area product are necessary for estimating OOB emissions. The FireCCI51 product (250 m), which features an advanced burned area algorithm, enhances the detection of fire locations and minimizes the omission of small fires12. Compared to the MCD64A1 product (500 m), the FireCCI51 algorithm offers higher accuracy in detecting burned areas, particularly for fires smaller than 100 hectares, thereby reducing the rate of missed detections13. In addition, FireCCI51 has the same temporal resolution as mainstream BA products (e.g. MCD64A1). Therefore, it is possible to develop emission inventories with high temporal and spatial resolution based on this product. Besides the burned area (BA) parameter, factors such as biomass fuel, combustion efficiency, and emission factors also significantly impact emission accuracy. Previous studies often used the average aboveground biomass density of biomass fuels14. For instance, Wu developed a 1km resolution OOB emission inventory of central and eastern China in which biomass fuels were averaged across provinces15. It is too coarse to account for biomass spatial variability and is suitable for large-scale study areas. Consequently, this study employed a high-resolution forest biomass product, developed by integrating various remote sensing observations and extensive field measurements16. It also utilized a high-resolution grassland biomass estimate derived from empirical equations to serve as biofuel factors. Similarly, the combustion efficiency (CE) was usually set to be consistent, without considering the differences between different vegetation types17. Thus, the CE of forest and grassland was dynamically calculated based on tree cover (TC) in this study. For cropland, the combustion efficiency of various crops was compiled. Additionally, the crop residue-to-production ratio (N) and dry matter fraction (D) of different crops were gathered to aid in estimating the effective combustion of biomass residue. Emission factors represent a critical parameter for estimating pollutants, making the selection of appropriate factors essential. The emission factors were collected systematically for various biomass types from an extensive review of published research. Finally, the appropriate emission factors for the study area were selected based on their alignment with the local environmental conditions. Heilongjiang Province, located in Northeast China (Fig. 1), is a major grain-producing region, accounting for 11% of China’s total agricultural output18. The open burning of crop residues releases substantial air pollutants, constituting a significant source of OBB emissions19,20. Additionally, Heilongjiang has a forest coverage rate of 43.78%, with a total forest area of 199,046 hectares. The province leads the nation in forest area, timber reserves, and timber production21,22. Emissions from frequent forest fires, second only to crop residue burning, cannot be ignored23. However, previous studies on OBB in Northeast China have produced emission inventories with limited temporal and spatial resolution. Additionally, many studies have focused on emissions from specific land-use categories24–26. There is a lack of systematic, high-resolution temporal and spatial emission inventory research for this region. Therefore, it is crucial to systematically develop long-term, high-resolution air pollutant emission inventories. In this study, we developed a high-resolution (250 m), multi-year (2001–2020) OBB emission inventory using monthly scale satellite data, dynamically calculated combustion efficiency factors from the tree cover dataset, high-resolution gridded biomass data, the latest pollutant emission factors from the NEC collected in the literature, and various types of survey data. It is worth noting that this study focuses on the spatial and temporal emission characteristics of PM2.5, which is a key indicator of air quality. With its small particle size, high concentration of toxic substances, and long residence time and transport distance in the atmosphere, PM2.5 significantly impacts air quality and even endangers human health27,28. This detailed multi-year inventory can be utilized in atmospheric transport models, aiding in the formulation of effective pollution mitigation strategies.Fig.1Land use type in Heilongjiang Province. Land use type in Heilongjiang Province. Materials and methods Study area Figure 1 depicts a detailed representation of the administrative subdivisions, the geographical positioning of Heilongjiang Province within China, and the spectrum of land use types in the province. Situated in northeastern China, Heilongjiang Province extends from longitude 121°11′ to 135°05′E and latitude 43°26′ to 53°33′N, and is comprised of 12 prefecture-level cities and one region (Daxinganling). In this study, we employed CLCD land use data, focusing on emissions from farmland, forests, and grasslands. Methods Typically, OBB calculations rely on factors such as the burned area, biomass fuel load, combustion efficiency, and emission factors29, thus:1\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\text{Emissions}} = \sum\limits_{i}^{n} {BA \times F \times CE \times EF} i,j$$\end{document}where i represents the land type, j represents each pollutant; BA is the burned area (m2); F is the available biomass fuel (kg m2); CE is the combustion efficiency, expressed as the ratio of actual biomass fuel burned to the total biomass burned; EF is the emission factor, which expresses the pollutant emissions corresponding to each kilogram of dry matter burned (g kg-1). Materials Burned area The burned area data were obtained using the FireCCI51 product released by ESA (http://cci.esa.int/data). This product uses MODIS Red (620–670 nm) and NIR (841–876 nm) solar surface reflectance data at a spatial resolution of 250 m, provided by the MODO9GQ products. FireCCI51 employs an updated burned area (BA) algorithm to synthesize burn trace data in a two-stage process. Initially, restrictive conditions filter high-probability burn pixels (seeds) from thermal anomaly information. Subsequently, contextual region growing is applied from these “seeds” to fully detect burn patches, thereby reducing omission errors30. The FireCCI51 product, unlike other BA hybrid algorithms, uses clustering-based thresholds during the seed and area growth stages, enhancing the identification of burned-unburned reflectance thresholds. It operates independently of land-use classifications, offering improved spatial and temporal resolution, and provides superior capability in detecting potentially burned patches compared to the preceding product and MCD64A1. Currently, many studies utilize Firecci51’s burned area data for research and have achieved positive outcomes31,32. In this study, we utilized the 250 m resolution pixel BA product provided by FireCCI51 to obtain monthly burned area data from 2001 to 2020. We set the confidence level to 80 using the CONFIDENCE band, ensuring that the burned pixels have a fairly high level of confidence. Biomass fuels Forests and grassland Forest biomass was derived from the 1 km resolution aboveground biomass carbon dataset of Chinese forests for the past 20 years (https://doi.org/10.6084/m9.figshare.21931161.v1). The dataset was calibrated against the GlobBiomass 2010 dataset using data from 2,444 sample plots, and the results were used as a baseline. Combined with optical and microwave remote sensing data, nonlinear regression estimates of biomass carbon were established for a long time series16. The above-ground biomass density of grassland was correlated with the annual maximum normalized difference vegetation index (AMNDVI), and the relationship can be expressed using the following power function:33,34:2\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\text{AGB}} = 179.71{\text{AMNDVI}}^{{{1}{\text{.6228}}}}$$\end{document}where AGB is the estimated aboveground biomass density of grassland (g C m-2) and AMNDVI is the maximum NDVI during the year, the formula is simple to calculate and stable. Specifically, the AMNDVI value of each pixel was obtained using 1 km resolution MOD13A monthly scale data on GEE, and then the above-ground biomass of the grassland was calculated using the formula. All data were ultimately maintained at the same resolution and projected coordinate system as the FireCCI51 product. Cropland Crop residues refer to the dry matter produced from crops such as rice, wheat, corn, soybean, potato, rapeseed, beet, tobacco, and others, which are burned on the cropland as fertilizer for the next year’s cultivation. The burning of waste from major crops is calculated from crop yields, straw production rates, and burning rates as follows:3\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$A_{i} = P_{i} \times N \times R \times D \times CEi,j$$\end{document}where i represents crop type, j represents each pollutant, Ai is the amount of crop residue burned in the field (kg year-1); Pi is the annual crop yield (kg year-1), N is the straw yield ratio for each crop type; R is the field straw burning rate; D is the dry matter content of each crop residue; and CE is the combustion efficiency of each crop residue type. In order to accurately quantify the extent of crop straw burning in Heilongjiang, which necessitates a slightly distinct approach compared to forests and grasslands when calculating emissions, we acquired data on annual crop production from the Heilongjiang Agricultural Statistical Yearbook (http://tjj.hlj.gov.cn/tjj/c106782/common_zfxxgk.shtml). For each crop type, data for N, D and CE were collected from published literature and local statistics (Table 1). The straw burning rate R is important for OBB emission estimation because it can reflect the implementation of regional and national control policies. Straw burning rate in Heilongjiang was about 30% before 2009 and about 22.6% thereafter35,36.Table 1Crop residue-to-production ratio (N), dry matter fraction (D), and combustion efficiency (CE) used in this study.NaDCERice0.90.890.93dWheat1.10.890.92dCorn1.20.870.92dSoybean1.60.910.68b,cPotato0.50.450.68b,cRapeseed1.50.830.9b,cBeet0.080.450.9b,cTobacco1.90.830.9b,caBi et al.37, bHe et al.38, cPeng et al.35, dZhou et al.39. Table 1Crop residue-to-production ratio (N), dry matter fraction (D), and combustion efficiency (CE) used in this study.NaDCERice0.90.890.93dWheat1.10.890.92dCorn1.20.870.92dSoybean1.60.910.68b,cPotato0.50.450.68b,cRapeseed1.50.830.9b,cBeet0.080.450.9b,cTobacco1.90.830.9b,caBi et al.37, bHe et al.38, cPeng et al.35, dZhou et al.39. Crop residue-to-production ratio (N), dry matter fraction (D), and combustion efficiency (CE) used in this study. aBi et al.37, bHe et al.38, cPeng et al.35, dZhou et al.39. The land-use product was derived from the 30 m resolution China CLCD dataset (https://zenodo.org/records/4417810), which is a collection of long time-series land-use data for China from 1990 to the present40. Specifically, the extent of cropland was first determined according to the CLCD product, and then we allocated the emissions from crop residue burning to each grid cell by overlaying the fire counts for each crop type. Agricultural fire frequency is the ratio of the number of fires in the month to the total number of fires in the year within the farmland boundaries, as follows:4\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_{i} = \frac{{FC_{i} }}{{FC_{k} }} \times E_{k}$$\end{document} Ei is the result of monthly pollutant allocation, FCi is the number of fires in the month, FCk is the total number of fires in the corresponding year, and Ek is the total amount of each pollutant emitted in the corresponding year. Combustion efficiency The combustion factor (CF) represents the proportion of biomass actually consumed by a fire41. It would have been set as a constant in previous studies, but this could have biased the emission estimates. Tree cover (TC) can be used to calculate CF for each fuel42,43. Tree cover data were calculated by using the annual-scale vegetation continuous fields (VCF) product, which corresponds exactly to TC. The combustion efficiencies (CEs) were set based on tree cover percentages as follows: For areas with more than 60% tree cover, the CEs for woody and herbaceous cover were set to 0.3 and 0.9, respectively. For areas with less than 40% tree cover, the CEs were set to 0 for woody cover and 0.98 for herbaceous cover. In areas with 40–60% tree cover, the CE for woody fuels was set to 0.3. for herbaceous cover was calculated with reference to the following equations:5\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\text{CE}}_{{\text{s}}} = {\text{e}}^{{ - 0.13 \times {\text{TB}}}}$$\end{document} TB is the percentage of tree coverage for the corresponding pixel. Emission factors Emission factors represent the amount of pollutants released during combustion, signalling trace gases and aerosols accompanying the combustion process44. In order to guarantee the precision of the biomass combustion emission inventory, it is of the utmost importance to select the appropriate emission factors during the study. We compared EF data for different biomass types from a large number of published research papers. EF experimental data from local measurements in the literature will be prioritised for specific vegetation types, particularly those closely related to the environmental conditions of the study area.45,46. The detailed emission factors appropriate for each vegetation type in the Northeast have been finalized in Table 247,48.Table 2Emission factors for different vegetation types.VegetationLCTBCCH4COCO2NH3NMVOCNoxOCPM10PM2.5SO2Evergreen coniferous forestforest0.2(0.2)6(3.1)118(45)1514(121)3.5(2.3)28(8.7)1.8(0.7)7.8(4.8)18.57(-)13(5.9)1(0.3)Evergreen broadleaved forestforest0.5(0.3)5.1(2.0)92(27)1663(58)0.8(1.2)24(-)2.6(1.4)4.7(2.7)13.86(-)9.7(3.5)0.5(0.2)Deciduous coniferous forestforest0.2(0.2)6(3.1)118(45)1514(121)3.5(2.3)28(8.7)3(0.7)7.8(4.8)19.43(-)13.6(5.9)1(0.3)Deciduous broadleaved forestforest0.6(0.2)5(0.9)102(19)1630(37)1.5(0.4)11(8.7)1.3(0.6)9.2(4.8)18.57(-)13(5.6)1(0.3)Mixed Forestforest0.6(0.2)5(0.9)102(19)1630(37)0.5(0.4)14(8.7)1.3(0.6)2.6(1.2)7.71(-)5.4(3.4)1(0.3)Grasslandgrassland0.4(0.2)1.5(0.9)59(17)1692(38)1.2(0.4)9.3(2.3)2.8(0.8)6.6(1.2)13.29(-)9.3(3.4)0.5(0.3)Ricecropland0.52(0.21)3.9(0.5)64.2(4.9)791(12.5)0.53(0.16)6.05(1.0)1.81(0.39)1.96(0.67)10.81(-)7.57(0.03)0.147(0.1)Wheatcropland0.49(0.17)3.4(0.5)60(23)1557(85.8)0.37(0.16)7.5(1.15)3.3(1.0)2.7(0.67)10.86(-)7.6(1.9)0.85(0.31)Corncropland1.55(0.86)4.4(0.5)53(8)1261(59.9)0.68(0.3)10(2.5)4.3(1.0)6.88(2.8)14.86(-)10.4(1.3)0.44(0.01)Soybeancropland0.23(0.1)3.9(0.5)32.3(1.1)1445(46)0.53(0.15)8.64(2.38)1.12(0.04)1.08(0.05)2.63(-)1.84(1.5)0.25(0.02)Potatocropland0.41(0.18)3.9(0.5)55.13(15.1)1445(46)0.53(0.15)8.64(1.36)2.11(0.1)2.03(1.05)8.23(-)5.76(1.44)0.25(0.02)Rapeseedcropland0.23(0.1)3.9(0.5)34.3(1.69)1445(46)0.53(0.15)8.64(1.36)1.12(1.56)1.08(0.5)4.74(-)3.32(1.48)0.25(0.02)Beetcropland0.41(0.18)3.9(0.5)55.13(15.1)1445(46)0.53(0.15)8.64(1.36)2.11(0.1)2.03(1.05)8.23(-)5.76(1.44)0.25(0.02)Tobaccocropland0.41(0.18)3.9(0.5)55.13(15.1)1445(46)0.53(0.15)8.64(1.36)2.11(0.1)2.03(1.05)8.23(-)5.76(1.44)0.25(0.02) Table 2Emission factors for different vegetation types.VegetationLCTBCCH4COCO2NH3NMVOCNoxOCPM10PM2.5SO2Evergreen coniferous forestforest0.2(0.2)6(3.1)118(45)1514(121)3.5(2.3)28(8.7)1.8(0.7)7.8(4.8)18.57(-)13(5.9)1(0.3)Evergreen broadleaved forestforest0.5(0.3)5.1(2.0)92(27)1663(58)0.8(1.2)24(-)2.6(1.4)4.7(2.7)13.86(-)9.7(3.5)0.5(0.2)Deciduous coniferous forestforest0.2(0.2)6(3.1)118(45)1514(121)3.5(2.3)28(8.7)3(0.7)7.8(4.8)19.43(-)13.6(5.9)1(0.3)Deciduous broadleaved forestforest0.6(0.2)5(0.9)102(19)1630(37)1.5(0.4)11(8.7)1.3(0.6)9.2(4.8)18.57(-)13(5.6)1(0.3)Mixed Forestforest0.6(0.2)5(0.9)102(19)1630(37)0.5(0.4)14(8.7)1.3(0.6)2.6(1.2)7.71(-)5.4(3.4)1(0.3)Grasslandgrassland0.4(0.2)1.5(0.9)59(17)1692(38)1.2(0.4)9.3(2.3)2.8(0.8)6.6(1.2)13.29(-)9.3(3.4)0.5(0.3)Ricecropland0.52(0.21)3.9(0.5)64.2(4.9)791(12.5)0.53(0.16)6.05(1.0)1.81(0.39)1.96(0.67)10.81(-)7.57(0.03)0.147(0.1)Wheatcropland0.49(0.17)3.4(0.5)60(23)1557(85.8)0.37(0.16)7.5(1.15)3.3(1.0)2.7(0.67)10.86(-)7.6(1.9)0.85(0.31)Corncropland1.55(0.86)4.4(0.5)53(8)1261(59.9)0.68(0.3)10(2.5)4.3(1.0)6.88(2.8)14.86(-)10.4(1.3)0.44(0.01)Soybeancropland0.23(0.1)3.9(0.5)32.3(1.1)1445(46)0.53(0.15)8.64(2.38)1.12(0.04)1.08(0.05)2.63(-)1.84(1.5)0.25(0.02)Potatocropland0.41(0.18)3.9(0.5)55.13(15.1)1445(46)0.53(0.15)8.64(1.36)2.11(0.1)2.03(1.05)8.23(-)5.76(1.44)0.25(0.02)Rapeseedcropland0.23(0.1)3.9(0.5)34.3(1.69)1445(46)0.53(0.15)8.64(1.36)1.12(1.56)1.08(0.5)4.74(-)3.32(1.48)0.25(0.02)Beetcropland0.41(0.18)3.9(0.5)55.13(15.1)1445(46)0.53(0.15)8.64(1.36)2.11(0.1)2.03(1.05)8.23(-)5.76(1.44)0.25(0.02)Tobaccocropland0.41(0.18)3.9(0.5)55.13(15.1)1445(46)0.53(0.15)8.64(1.36)2.11(0.1)2.03(1.05)8.23(-)5.76(1.44)0.25(0.02) Emission factors for different vegetation types. Method for uncertainty analysis The process of establishing an emission inventory of air pollution sources is susceptible to random errors in data collection and the selection of emission factors. These errors may introduce uncertainty into the results of the inventory estimates. The uncertainty can be estimated using Monte Carlo simulation methods49–51. Monte Carlo simulation methods, based on probabilistic statistical theory, can analyze various uncertainty problems. This method assumes familiarity with the probability distribution function of the random variable. It utilizes random sampling techniques to create numerous sets of input variables, each set being simulated to yield a large quantity of outputs. The outputs’ statistical properties are then examined, including maximum range, mean, variance, and other pertinent statistical measures. Results and discussion Inventory of OOB emissions in Heilongjiang Estimating the average annual open biomass burning emissions for each pollutant at a 250 m resolution, the average annual amounts for Heilongjiang Province from 2001 to 2020 were found to be as follows:BC:12.26 Gg year-1, CH4:63.70 Gg year-1, CO:931.60 Gg year-1, CO2:17,203.35 Gg year-1, NH3:16.09 Gg year-1, NMVOC:171.25 Gg year-1, Nox:39.83 Gg year-1, OC:71.22 Gg year-1, PM10:184.33 Gg year-1, PM2.5:129.16 Gg year-1, SO2:6.35 Gg year-1 respectively. The top pollutant emitted is CO2, which was strongly associated with the greenhouse effect. The average annual emissions of 11 pollutants were shown in Table 3, and the corresponding contribution were exhibited in Fig. 2. Crop residue burning and forest fires were the main sources of pollutant emissions, and the cumulative contribution of various pollutant emissions was in the range of 40–95%. Crop residue burning was the largest source of biomass combustion emissions, with the percentage of each pollutant emitted being BC (95.5%), CH4 (75.5%), CO (67%), CO2 (76.8%) NH3 (43.6%), NMVOC (57.4%), Nox (88%), OC (71.5%), PM10 (73.9%) PM2.5 (73.8%), SO2 (58.6%). Forest fires are also noteworthy, contributing significantly to emissions of pollutants such as NH3 (56.2%), NMVOC (42.2%), and SO2 (40.9%), which are comparable to emissions from crop residue burning. In grassland burning, the pollutants with the highest emission shares were CO2, followed by SO2 and Nox.Table 3The emissions of open biomass burning and the annual average emissions of each vegetation type, 2001 to 2020. (Gg year-1).BCCH4COCO2NH3NMVOCNoxOCPM10PM2.5SO220016.2953.38868.4914,441.4819.27178.1425.2757.90149.65104.986.4620026.7037.07516.7910,582.018.3994.9922.3837.3498.5369.033.5420036.6768.831169.0818,444.1928.82254.2329.4478.75195.90137.499.11200410.20138.832500.3436,286.0266.61562.4352.91166.99408.61286.9120.1820058.4444.03591.6812,569.828.88106.1027.3743.52115.2680.723.9520068.9747.31658.0213,175.3810.24117.2229.2348.04128.1789.784.3520079.6044.32606.5512,055.747.7997.7229.8645.18124.5287.193.64200811.8163.73926.6417,288.4315.54167.4038.8467.86179.95126.096.1820099.6560.19931.4016,122.9918.28179.7633.7267.06173.80121.846.62201011.5865.13986.1017,406.9018.09184.2538.8172.93189.92133.106.80201112.8261.13872.9516,486.4513.54152.4840.4967.84178.29124.895.73201213.7354.50722.5514,710.468.22112.6240.8060.21160.42112.304.32201315.3962.28847.7016,645.5910.99138.4845.9971.65186.91130.875.28201416.1263.99860.2317,210.6510.69139.0047.9273.40191.41134.015.33201517.4673.971036.3819,854.6615.47180.1853.0787.99223.92156.826.86201616.1567.08917.8318,037.2112.50153.6348.6377.12199.80139.915.83201715.7371.231000.5519,157.4715.07173.7148.6380.50209.15146.496.50201816.2564.78859.3117,569.3610.74140.8348.2673.88191.08133.785.39201916.2467.91914.7818,521.4112.39154.8348.9477.03198.37138.905.86202015.3964.29845.0717,500.8010.30136.9146.0569.13182.91128.055.11Forest0.5215.52305.123914.779.0572.44.6620.1747.9433.622.6Grassland0.020.072.7879.860.020.440.130.120.360.250.03Cropland11.7248.11623.7213,208.727.0298.4135.0450.93136.1395.293.72Mean12.2663.7931.6217,203.3516.09171.2539.8371.22184.33129.166.35Fig. 2Cumulative percentage of average annual OBB emissions for each land type in Heilongjiang during the period 2001–2020. Table 3The emissions of open biomass burning and the annual average emissions of each vegetation type, 2001 to 2020. (Gg year-1).BCCH4COCO2NH3NMVOCNoxOCPM10PM2.5SO220016.2953.38868.4914,441.4819.27178.1425.2757.90149.65104.986.4620026.7037.07516.7910,582.018.3994.9922.3837.3498.5369.033.5420036.6768.831169.0818,444.1928.82254.2329.4478.75195.90137.499.11200410.20138.832500.3436,286.0266.61562.4352.91166.99408.61286.9120.1820058.4444.03591.6812,569.828.88106.1027.3743.52115.2680.723.9520068.9747.31658.0213,175.3810.24117.2229.2348.04128.1789.784.3520079.6044.32606.5512,055.747.7997.7229.8645.18124.5287.193.64200811.8163.73926.6417,288.4315.54167.4038.8467.86179.95126.096.1820099.6560.19931.4016,122.9918.28179.7633.7267.06173.80121.846.62201011.5865.13986.1017,406.9018.09184.2538.8172.93189.92133.106.80201112.8261.13872.9516,486.4513.54152.4840.4967.84178.29124.895.73201213.7354.50722.5514,710.468.22112.6240.8060.21160.42112.304.32201315.3962.28847.7016,645.5910.99138.4845.9971.65186.91130.875.28201416.1263.99860.2317,210.6510.69139.0047.9273.40191.41134.015.33201517.4673.971036.3819,854.6615.47180.1853.0787.99223.92156.826.86201616.1567.08917.8318,037.2112.50153.6348.6377.12199.80139.915.83201715.7371.231000.5519,157.4715.07173.7148.6380.50209.15146.496.50201816.2564.78859.3117,569.3610.74140.8348.2673.88191.08133.785.39201916.2467.91914.7818,521.4112.39154.8348.9477.03198.37138.905.86202015.3964.29845.0717,500.8010.30136.9146.0569.13182.91128.055.11Forest0.5215.52305.123914.779.0572.44.6620.1747.9433.622.6Grassland0.020.072.7879.860.020.440.130.120.360.250.03Cropland11.7248.11623.7213,208.727.0298.4135.0450.93136.1395.293.72Mean12.2663.7931.6217,203.3516.09171.2539.8371.22184.33129.166.35 The emissions of open biomass burning and the annual average emissions of each vegetation type, 2001 to 2020. (Gg year-1). Cumulative percentage of average annual OBB emissions for each land type in Heilongjiang during the period 2001–2020. PM2.5 spatial pattern and variation of emissions PM2.5 emissions were higher in western, southwestern, and northeastern Heilongjiang. These areas are the primary agricultural areas in Heilongjiang, having advantageous climatic conditions and fertile soils. The burning of significant amount of crop residue has become a major source of OOB emissions in the region. Specifically, the prefecture-level cities of Qiqihar, Daqing, Shuangyashan, and Jiamusi were focal points. Within the average annual fire occurrence range, PM2.5 emissions exceeded 900 kg a-1 per 250 m resolution grid. PM2.5 emissions in the central and northern regions were medium, and the specific prefecture-level cities were Suihua City, southern Heihe City and other places, with a range of 300–500 kg a-1. Low emissions were in the south and northwest. Probably due to strict straw burning policies in these regions. Specific prefecture-level cities were Harbin City, Mudanjiang City and other places, with emissions less than 100 kg a-1 (Fig. 3a).Fig. 3Spatial distribution of average annual OBB PM2.5 emissions in Heilongjiang during the period 2001–2020: (a) total; (b) cropland; (c) forest; (d) grassland; Spatial distribution of average annual OBB PM2.5 emissions in Heilongjiang during the period 2001–2020: (a) total; (b) cropland; (c) forest; (d) grassland; Among various vegetation types, emissions from cropland fires were widely distributed and contributed the most to PM2.5, accounting for about 73%. Emissions at most locations were greater than 300 kg a-1 (Fig. 3b). Forest fires were mainly concentrated in northern Heilongjiang (Fig. 3c), which region was Daxinganling forest area of Heihe City. The higher biomass density causes more PM2.5 to be emitted when fires occur, with emissions ranging from about 100 to 300 kg a-1. The emission distribution of grassland burning was sporadic (Fig. 3d), which is due to the lower biomass density of grasslands resulting in lower emissions, less than 100 kg a-1. PM2.5 temporal trends According to the emission factors of each vegetation type, the annual emission changes of 11 types of air pollutants in Heilongjiang were shown in Fig. 4. PM2.5 emissions exhibited a sharp increase around 2004, followed by a brief decrease from 2005 to 2008, and subsequently stabilized over the long term (2008–2020). This was due to the fact that the emissions of various pollutants have been effectively controlled after the implementation of the policy of strictly limiting straw burning in China52–55. Compared with studies by other scholars, the average annual emission of PM2.5 in Heilongjiang from 2001 to 2017 was 116.9Gg year-148. During the same period of our study, the average annual PM2.5 emission in Heilongjiang was 128.38Gg year-1, and the total emission value was slightly higher (similar results were found for other pollutant emissions). The higher OOB emission estimates were likely attributed to the FireCCI51 product’s enhanced spatial resolution and improved detection of small fires.Fig. 4Yearly OBB emissions in Heilongjiang during the period 2001–2020. Yearly OBB emissions in Heilongjiang during the period 2001–2020. Monthly variations in PM2.5 showed distinct patterns (Fig. 5). During the study period, peak PM2.5 emissions occurred in April and October (total = 78.13 Gg), comprising 60.5% of the total year. Emissions from crop residue burning were concentrated in spring and autumn, accounting for 48.76% and 47.58% of the year, respectively (Fig. 6a). Emissions from the forest fires were mainly in autumn, accounting for 62.78% of the year, followed by spring with a share of 36.24% (Fig. 6b). Emissions from the grassland were also concentrated in spring and autumn, accounting for 54.35% and 41.67%, respectively (Fig. 6c). Overall emissions were distinctly seasonal, with the highest emissions occurring in autumn (total = 66.55 Gg), accounting for 51.5% of the year, followed by spring (total = 58.78 Gg), and the lowest emissions occurring in winter and summer (total = 3.8 Gg). Crop residue burning and forest fires remain the dominant types of PM2.5 emissions. Crop residue burning predominantly occurred during the post-harvest period and the pre-sowing season. The former was driven by the unscientific and inefficient management of the substantial straw left after harvest, leading to its disposal through open burning. The latter was associated with land reclamation practices, where on-site burning not only enhances soil fertility for subsequent planting but also reduces the cost of disposal. In Heilongjiang, October and November are the harvest seasons for corn and rice, during which a significant amount of straw burning leads to high emissions. In forested areas, the dry climate conditions during spring and autumn exacerbated the risk of wildfires, which were further exacerbated by the increase in sacrificial activities during the Qingming Festival (late March to early April), leading to elevated PM2.5 emissions56,57.Fig. 5Monthly OBB PM2.5 emissions in Heilongjiang during the period 2001–2020.Fig. 6Seasonally percentage of OBB PM2.5 emissions in Heilongjiang during the period 2001–2020: (a) cropland; (b) forest; (c) grassland. Monthly OBB PM2.5 emissions in Heilongjiang during the period 2001–2020. Seasonally percentage of OBB PM2.5 emissions in Heilongjiang during the period 2001–2020: (a) cropland; (b) forest; (c) grassland. Uncertainties The uncertainty of emission inventories primarily arises from variability of parameters used in the estimation process. Emissions from forest fires and grasslands were calculated by multiplying factors such as burned area, aboveground biomass, combustion completeness (CE), and emission factors (EF). And cropland type was estimated including crop yield, grass-to-grain ratio, dry matter fraction, burn ratio, burn efficiency, and crop-specific emission factors. The uncertainty of the forest and grassland biomass data is 50% of the mean biomass, the crop yield is 5%, the CE is 5–30%, the proportion of straw burned is 30%, and the dry matter fraction is 10%48,58. By integrating the above parameters, we performed 200,000 Monte Carlo simulations to estimate the range of annual emissions of each pollutant with 95% confidence intervals (Table 4). The total range of PM2.5 emissions 2001–2020 was estimated to be 57.56–206.20 Gg year-1. Updating emission factors, burned areas, and other parameters can enhance the accuracy of biomass combustion emission estimation. Future research should focus on enhancing the spatial and temporal resolution. Additionally, integrating emissions from different vegetation types across different temporal scales can further enhance the accuracy of emission inventory estimates.Table 4Estimation of uncertainty in emissions from open biomass combustion with 95% confidence interval during the period 2001–2020 (Gg year-1).yearBCCH4COCO2NH3NMVOCNOXOCPM10PM25so22001(1.13, 12.8)(16.55, 95.13)(230.9, 1526.61)(5529.11, 21,652.48)(-1.4, 43.95)(35.57, 316.9)(10.35, 41.91)(4.51, 122.83)(39.2, 274.46)(28.36, 189.6)(1.32, 11.38)2002(1.69, 13.09)(19.95, 56.43)(255.1, 802.02)(6109.58, 14,793.12)(1.24, 17.2)(38.84, 155.34)(11.83, 35.03)(10.11, 71.33)(47.21, 157.66)(33.16, 109.75)(1.49, 5.66)2003(0.57, 14.35)(14.25, 131.03)(209.7, 2157.56)(5306.31, 28,637.14)(-3.77, 67.71)(34.23, 465.64)(9.63, 51.38)(0.17, 173.3)(32.58, 382.6)(24.15, 262.7)(1.25, 16.51)2004(-0.2, 21.93)(14.91, 251.1)(259.14, 4260.7)(6736.69, 52,486.42)(-9.53, 140.45)(43.56, 933.34)(12.17, 88.47)(-8.48, 339.07)(34.17, 738.04)(28.03, 512.77)(1.62, 33.32)2005(2.23, 16.23)(25.4, 63.48)(318.84, 860.16)(7669.38, 16,982.88)(2.02, 16.32)(48.86, 161.61)(14.88, 41.8)(13.81, 80.14)(59.89, 174.08)(42.22, 121.38)(1.9, 5.86)2006(2.31, 17.67)(26.31, 71.33)(336.85, 1013.34)(7774.15, 18,331.91)(1.83, 20.52)(50.6, 188.61)(15.44, 45.4)(13.59, 91.25)(64.03, 201.49)(45.01, 141.11)(1.89, 6.86)2007(2.5, 18.51)(26.88, 63.6)(353.51, 880.2)(7616.27, 16,597.57)(2.53, 14.06)(50.2, 150.01)(16.36, 46.39)(15.95, 82.02)(70.95, 184.49)(49.72, 128.01)(1.83, 5.53)2008(2.8, 23.13)(32.05, 94.01)(421.01, 1352.82)(9292.9, 23,410.4)(2.01, 28.37)(61.69, 252.65)(19.59, 60.27)(16.96, 123.83)(83.07, 270.61)(58.41, 189.11)(2.28, 9.19)2009(1.97, 20)(24.36, 97.01)(332.86, 1479.2)(7394.24, 22,894.68)(0.25, 37.08)(49.46, 291.36)(15.49, 54.52)(11.48, 130.23)(64.28, 286.86)(45.53, 198.3)(1.85, 10.57)2010(2.35, 23.48)(28.96, 101.83)(393.73, 1536.23)(8542.94, 24,571.87)(0.97, 35.85)(56.47, 294.74)(18.35, 62.3)(14.74, 139.8)(78.56, 303.63)(55.73, 210.45)(2.13, 10.71)2011(2.79, 26.32)(32.43, 97.35)(436.59, 1406.04)(9373.15, 24,242.57)(2.02, 29.06)(61.91, 260.57)(20.65, 66.23)(18.16, 135.18)(88.54, 290.52)(62.82, 202.98)(2.36, 9.53)2012(3.29, 27.94)(34.98, 78.33)(458.47, 1038.91)(9635.87, 20,716.04)(3.54, 14.23)(64.12, 174.55)(22.4, 65.11)(22.22, 111.59)(98.36, 237.2)(68.48, 165.73)(2.45, 6.51)2013(3.09, 31.31)(36.85, 92.29)(484.4, 1258.44)(10,045.62, 23,735.65)(3.3, 19.96)(68.5, 220.49)(23.86, 74.63)(23.34, 135.36)(104.61, 283.89)(73.64, 197.26)(2.59, 8.13)2014(3.54, 32.76)(38.73, 92.57)(511.53, 1251.09)(10,684.68, 24,260.47)(3.73, 19.01)(71.91, 218.02)(25.13, 77.55)(24.37, 137.74)(109.24, 284.32)(76.94, 198.86)(2.76, 8.03)2015(3.41, 36.68)(39.68, 112.7)(526.72, 1584.89)(11,021.37, 28,749.68)(2.89, 29.55)(74.8, 293.38)(26.39, 86.9)(24.6, 168.86)(114.03, 349.04)(80.27, 244.73)(2.88, 10.74)2016(3.35, 33.11)(39.16, 99.14)(517.26, 1353.74)(10,831.99, 25,670.77)(3.52, 22.54)(72.54, 240.11)(25.25, 78.71)(24.35, 145.5)(110.49, 303.15)(78.01, 213.08)(2.77, 8.88)2017(3.39, 32.2)(38.41, 106.63)(503.33, 1502.98)(10,803.89, 27,146.19)(2.94, 28)(73.1, 272.79)(24.47, 79.01)(22.61, 152.45)(105.72, 321.09)(74.14, 224.61)(2.73, 9.98)2018(3.43, 33.57)(39.24, 94.6)(515.45, 1261.78)(10,903.1, 24,986.9)(3.81, 19.23)(72.75, 221.34)(25.14, 78.42)(25.01, 138.44)(109.67, 287.74)(76.71, 202.33)(2.77, 8.19)2019(3.64, 33.63)(39.62, 104.04)(512.6, 1415.8)(11,159.16, 27,069.83)(3.35, 24.96)(75.04, 258.85)(25.06, 80.48)(23.94, 150.54)(108.52, 312.48)(75.8, 219.15)(2.84, 9.53)2020(3.52, 31.31)(40.82, 92.26)(524.24, 1218.33)(11,335.36, 24,406.23)(4.05, 18)(76.21, 211.84)(24.91, 73.28)(24.8, 127.34)(108.76, 271.78)(75.79, 192.11)(2.76, 7.7)mean(2.54, 25)(30.48, 99.74)(405.11, 1458.04)(8888.29, 24,567.14)(1.47, 32.3)(59.02, 279.11)(19.37, 64.39)(16.31, 137.84)(81.59, 295.76)(57.65, 206.2)(2.22, 10.14) Table 4Estimation of uncertainty in emissions from open biomass combustion with 95% confidence interval during the period 2001–2020 (Gg year-1).yearBCCH4COCO2NH3NMVOCNOXOCPM10PM25so22001(1.13, 12.8)(16.55, 95.13)(230.9, 1526.61)(5529.11, 21,652.48)(-1.4, 43.95)(35.57, 316.9)(10.35, 41.91)(4.51, 122.83)(39.2, 274.46)(28.36, 189.6)(1.32, 11.38)2002(1.69, 13.09)(19.95, 56.43)(255.1, 802.02)(6109.58, 14,793.12)(1.24, 17.2)(38.84, 155.34)(11.83, 35.03)(10.11, 71.33)(47.21, 157.66)(33.16, 109.75)(1.49, 5.66)2003(0.57, 14.35)(14.25, 131.03)(209.7, 2157.56)(5306.31, 28,637.14)(-3.77, 67.71)(34.23, 465.64)(9.63, 51.38)(0.17, 173.3)(32.58, 382.6)(24.15, 262.7)(1.25, 16.51)2004(-0.2, 21.93)(14.91, 251.1)(259.14, 4260.7)(6736.69, 52,486.42)(-9.53, 140.45)(43.56, 933.34)(12.17, 88.47)(-8.48, 339.07)(34.17, 738.04)(28.03, 512.77)(1.62, 33.32)2005(2.23, 16.23)(25.4, 63.48)(318.84, 860.16)(7669.38, 16,982.88)(2.02, 16.32)(48.86, 161.61)(14.88, 41.8)(13.81, 80.14)(59.89, 174.08)(42.22, 121.38)(1.9, 5.86)2006(2.31, 17.67)(26.31, 71.33)(336.85, 1013.34)(7774.15, 18,331.91)(1.83, 20.52)(50.6, 188.61)(15.44, 45.4)(13.59, 91.25)(64.03, 201.49)(45.01, 141.11)(1.89, 6.86)2007(2.5, 18.51)(26.88, 63.6)(353.51, 880.2)(7616.27, 16,597.57)(2.53, 14.06)(50.2, 150.01)(16.36, 46.39)(15.95, 82.02)(70.95, 184.49)(49.72, 128.01)(1.83, 5.53)2008(2.8, 23.13)(32.05, 94.01)(421.01, 1352.82)(9292.9, 23,410.4)(2.01, 28.37)(61.69, 252.65)(19.59, 60.27)(16.96, 123.83)(83.07, 270.61)(58.41, 189.11)(2.28, 9.19)2009(1.97, 20)(24.36, 97.01)(332.86, 1479.2)(7394.24, 22,894.68)(0.25, 37.08)(49.46, 291.36)(15.49, 54.52)(11.48, 130.23)(64.28, 286.86)(45.53, 198.3)(1.85, 10.57)2010(2.35, 23.48)(28.96, 101.83)(393.73, 1536.23)(8542.94, 24,571.87)(0.97, 35.85)(56.47, 294.74)(18.35, 62.3)(14.74, 139.8)(78.56, 303.63)(55.73, 210.45)(2.13, 10.71)2011(2.79, 26.32)(32.43, 97.35)(436.59, 1406.04)(9373.15, 24,242.57)(2.02, 29.06)(61.91, 260.57)(20.65, 66.23)(18.16, 135.18)(88.54, 290.52)(62.82, 202.98)(2.36, 9.53)2012(3.29, 27.94)(34.98, 78.33)(458.47, 1038.91)(9635.87, 20,716.04)(3.54, 14.23)(64.12, 174.55)(22.4, 65.11)(22.22, 111.59)(98.36, 237.2)(68.48, 165.73)(2.45, 6.51)2013(3.09, 31.31)(36.85, 92.29)(484.4, 1258.44)(10,045.62, 23,735.65)(3.3, 19.96)(68.5, 220.49)(23.86, 74.63)(23.34, 135.36)(104.61, 283.89)(73.64, 197.26)(2.59, 8.13)2014(3.54, 32.76)(38.73, 92.57)(511.53, 1251.09)(10,684.68, 24,260.47)(3.73, 19.01)(71.91, 218.02)(25.13, 77.55)(24.37, 137.74)(109.24, 284.32)(76.94, 198.86)(2.76, 8.03)2015(3.41, 36.68)(39.68, 112.7)(526.72, 1584.89)(11,021.37, 28,749.68)(2.89, 29.55)(74.8, 293.38)(26.39, 86.9)(24.6, 168.86)(114.03, 349.04)(80.27, 244.73)(2.88, 10.74)2016(3.35, 33.11)(39.16, 99.14)(517.26, 1353.74)(10,831.99, 25,670.77)(3.52, 22.54)(72.54, 240.11)(25.25, 78.71)(24.35, 145.5)(110.49, 303.15)(78.01, 213.08)(2.77, 8.88)2017(3.39, 32.2)(38.41, 106.63)(503.33, 1502.98)(10,803.89, 27,146.19)(2.94, 28)(73.1, 272.79)(24.47, 79.01)(22.61, 152.45)(105.72, 321.09)(74.14, 224.61)(2.73, 9.98)2018(3.43, 33.57)(39.24, 94.6)(515.45, 1261.78)(10,903.1, 24,986.9)(3.81, 19.23)(72.75, 221.34)(25.14, 78.42)(25.01, 138.44)(109.67, 287.74)(76.71, 202.33)(2.77, 8.19)2019(3.64, 33.63)(39.62, 104.04)(512.6, 1415.8)(11,159.16, 27,069.83)(3.35, 24.96)(75.04, 258.85)(25.06, 80.48)(23.94, 150.54)(108.52, 312.48)(75.8, 219.15)(2.84, 9.53)2020(3.52, 31.31)(40.82, 92.26)(524.24, 1218.33)(11,335.36, 24,406.23)(4.05, 18)(76.21, 211.84)(24.91, 73.28)(24.8, 127.34)(108.76, 271.78)(75.79, 192.11)(2.76, 7.7)mean(2.54, 25)(30.48, 99.74)(405.11, 1458.04)(8888.29, 24,567.14)(1.47, 32.3)(59.02, 279.11)(19.37, 64.39)(16.31, 137.84)(81.59, 295.76)(57.65, 206.2)(2.22, 10.14) Estimation of uncertainty in emissions from open biomass combustion with 95% confidence interval during the period 2001–2020 (Gg year-1). Conclusions In this study, we developed a finer-resolution, long-term emission inventory for open biomass burning in Heilongjiang Province by integrating FireCCI51 burned area data, biomass data for each class, dynamically calculated combustion efficiencies (CEs), and pollutant emission factors (EFs). In comparison to previous studies, our estimates exhibit a slight increase due to the utilization of higher-resolution burned area data, which offers greater advantages in detecting small fires than the MCD64A1 data. The average annual amounts of pollutants for Heilongjiang Province (2001–2020) were found to be as follows:BC:12.26 Gg year-1, CH4:63.70 Gg year-1, CO:931.60 Gg year-1, CO2:17,203.35 Gg year-1, NH3:16.09 Gg year-1, NMVOC:171.25 Gg year-1, Nox:39.83 Gg year-1, OC:71.22 Gg year-1, PM10:184.33 Gg year-1, PM2.5:129.16 Gg year-1, SO2:6.35 Gg year-1 respectively. The emissions of various pollutants exhibited a significant increase (around 2004), followed by a brief decrease (2005–2008) and ultimately a long-term stabilization (2008–2020). Taking PM2.5 as an example, emissions were high in the western, southwestern and northeastern parts of Heilongjiang, moderate in the central and northern parts, and low in the southern and northwestern parts. Crop residue burning and forest fires are the main factors for higher emissions. The emissions of PM2.5 from cropland exhibited seasonal concentration, with 48.76% and 47.58% occurring in spring and fall, respectively. Forest fire emissions were predominantly observed in the fall (62.78%) and to a lesser extent in the spring (36.24%). Similarly, grassland emissions displayed significant concentration in both spring (54.35%) and fall (41.67%). The highest emission season was fall (total = 66.55 Gg) with 51.5% of the year, followed by spring (total = 58.78 Gg). Despite the inherent uncertainties in the emission inventories stemming from errors in input parameters, the integration of various high-resolution remotely sensed data and updated local emission factors promises to enhance the precision of the inventories. Moreover, while the strict enforcement of policies to curb straw burning has led to effective control over the emissions of various pollutants, a subset of pollutant emissions continues to exhibit an increasing trend. Consequently, the adoption of more efficient straw utilization strategies and the intensification of regional OOB emissions monitoring are imperative.